🔮 Artificial intelligence: at work, in management, for humanity; exponential scientists; Facebook vs Google; mean scripture kiddies; WeChat++ #34
|Azeem Azhar||Nov 8, 2015|
Will AI strip us of jobs or give us more meaningful work? Differing views shared. Rating service providers makes us mean (so does religion). The amazing growth of science. Facebook & Google face off over AI. Vacation special.
Referred by a friend? Sign up
Dept of the near future
👷 Workplace activities may be automated, but entire occupations are unlikely to be. Automation will in general free humans up for more creative and social tasks. Dry, but well structured, report by McKinsey. HIGHLY RECOMMENDED.
🔬 There has been an exponential growth in science in the past 100 years. Amazingly, more than 90% of scientists who have ever lived are alive today. WORTH PONDERING - interesting cultural implications.
😮 Ritz Carlton service at McDonald’s prices: Uber and it’s ilk give us the power to rate providers - and turn as into horrible bosses. MUST READ
🏢 Corporations are getting smaller and bigger: the Internet reduces transaction costs allowing more ‘virtual corporations’ but also reduces the cost of managing mega-corps, argue EV reader Roy Bahat.
The future of management is algorithmic argues Esko Kilpi. WORTH READING raises questions above management education
💡 “Ad blocking at this stage is mostly prophylaxis. But it is also a demonstration of unprecedented individual agency in the online world.” Ad blockers and the future of the Internet by the brilliant Doc Searls. EXCELLENT
🌟 AI will mean “homo sapiens will be split into a handful of ‘gods’, and then the rest of us”. Excellent review of AI’s effect of the economy by Charles Arthur. (Quotes from a long BofA report on creative destruction from April.)
💊 The $1000 genome has arrived (and is getting exponentially cheaper, even more so than our familiar Moore’s Law CPU-trend).
Dept of AI battles
There is an epic battle brewing between Facebook and Google on AI. The stakes are high: the two top consumer internet companies. AI systems can provide for better user experience, so consumer intention will follow.
From a technical perspective, both have scads of training data. Google’s : an impressive long-term archive of the Net, including public social postings, and billions of email messages; together with real data about what people click on when they search for things.
Facebook’s data the deepest, richest social graph in history, together with all the posts and interaction-level data from 20% of humanity. They are releasing their AI assistant, M, increasingly widely. (M allows you to ask for ‘anything legal’)
The blue corner: Facebook
Facebook is starting to tackle one of Computer Science’s toughest problems: can you teach a computer to win at the combinatorially-complex Japanese game Go? (I’m told privately that Google’s Deepmind, based in London, is also tackling Go.)
Good interview with Yann Lecun, Facebook’s director of AI science.
The multi-coloured corner: Google
😃 Google outlines their Smart Reply system in Gmail which uses recurrent neural networks and Geoff Hinton’s “thought vectors’ to automatically compose suitable replies to a given email.
(Incidentally, Siri on the Apple Watch seems to have a similar capability when responding to text messages, I don’t know how they are doing this, and if any of our readers at Apple want to risk their careers enough, please drop me a mail to explain…)
💥 How machine learning works according to Google. Excellent accessible notes by Danny Sullivan.
One huge area for AI is in improving human-computer interfaces. Facebook’s M is a good example of this; as is Amy, the calendaring app, and Weave, an AI-based mobile search platform I am advising.
Viv is another interesting play, founded by the founders of Siri, it promises to out-Siri Siri suggests startup guru Robert Scoble, who has had chance to play with it. 💥 Scoble is seriously impressed with Viv’s capabilities relative to Siri or Google Now. But what kind of real advantage they can have without access to the scale, breadth and depth of data that Google, Apple or Facebook have?
😀 EV reader Samim continues his recurrent neural network rampage now generating short stories from images. (Rather fun.)
Fabulous short video: what an autonomous car sees. (Don’t know the source.)
Short morsels for dinner parties
The Saudis will set some type of record for environmental mismanagement. Their desert wheat industry is about to run out of water. Fascinating read. We touched on this issue in EV#27, as well, recommend that for a deeper dive.
✟ Religious children are meaner than their secular counterparts (particularly Christian and Muslim kids).
Gender and STEM occupation predict where you likely lie on the autism spectrum. (Data backs our assumptions.)
Lytro, the light-field camera innovators, are showing of a 3-D light-field camera for VR application.
🇨🇳 WeChat dominates Chinese daily life. Here is an exceptionally handy review of the product and innovation philosophy behind the app; and how people use it. GOOD READ
The first clinical trials using CRISPR to treat a disease will take place in 2017.
Power-law distributions are common in physics. We see them in social systems as well. Apparently, they appear in ecology too.
⁇ Is billion-dollar Faraday Future really a front for the Apple Car?
Exploring the microbiome: more than 95% of the viruses pruning the bacterial colonies on our skin are unidentified
Twitter isn’t working - probably because Twitter doesn’t get Twitter. Colourful & thoughtful essay by Thomas Baedkal, if you are remotely interested in Twitter & product management.
What you wrote
Tech is dematerialising the economy. We need less physical stuff to produce the same amount, writes EV reader, Diane Coyle.
I am on holiday right now, so doing my best to keep Exponential View at the quality you like. Don’t forget to drop me a note if you see anything of interest or recommend this to friends.
Thanks for all of those who responded to the survey last week.